Cumulative sleep deprivation commonly results from operational demands involving transmeridian flights, night missions, and sustained operations. Our recent studies of both acute total sleep deprivation (TSD) and cumulative partial sleep deprivation (PSD) demonstrate that as sleep loss escalates, the temporal profiles of declining alertness and performance degradation reflect the influence of both the circadian pacemaker and the homeostatic sleep drive, but the interaction between the two processes is nonlinear in ways not predicted by current models of alertness. We are conducting an experiemnt simulating deployments that produce severe TSD or cumulative PSD in order to: (1) specify the contributions (and interactions) of circadian and homeostatic processes to neurobehavioral functions during sleep loss (88 hr total sleep loss vs. 2 hr naps every 12 hr for 88 hrs; (2) evaluate the effectiveness of sustained, low dose 0l3 mg/kg) caffeine (versus placebo) as a countermeasure to the neurobehavioral devicits produced by sleep deprivation; and (3) assess the role of recovery sleep duration (7hr vs 14 hr) in normalization of waking functions following varying degrees of sleep deprivation.